$ RAILS_ENV=production ruby mybb2.rb loading existing groups... loading existing users... loading existing categories... loading existing posts... loading existing topics... creating users /var/www/discourse/vendor/bundle/ruby/2.0.0/gems/rack-mini-profiler-0.9.7/lib/patches/db/mysql2.rb:6: [BUG] Segmentation fault ruby 2.0.0p647 (2015-08-18 revision 51631) [x86_64-linux] [... about 3000 lines snipped ...] [NOTE] You may have encountered a bug in the Ruby interpreter or extension libraries. Bug reports are welcome. For details: http://www.ruby-lang.org/bugreport.html Aborted
I’m seeing the same on this end as well, not in the same place, but definitely the same segfault.
I had a similar problem with the phpBB3 importer.
Try changing the
If that doesn’t help try upgrading the mysql2 gem:
gem update mysql2
Oh cool, that looks like it works. Thanks!
cache_rows: true prevents the crash.
MyBB imports used to work until a few weeks ago. The crash is a regression and blocks not only imports, but also certain kinds of batch processing.
cache_rows altogether also prevents the crash. I did downgrade the mysql gem to v0.3.20 though.
I JUST saw that today … we got to raise this issue with the mysql2 gem…its crash central with cache_rows: true. Can you raise that with them together with the segfault?
This is a reasonably new issue, was working fine a few weeks ago. Now it is bust on pretty much every ruby version.
Never mind … will raise …
Thanks @sam. How can one revert to a working state of the gem and whatever else is involved? I only need to obtain a map of old_topic_id -> imported topic_id to create a redirect map file, and the importer itself was the easiest way to get that.
I can’t find the core dump and am in the middle of an import, can you do a gist of it?
Unless you are a paying customer, you are not obligated to our time. If you continue to make demands like this, either buy a hosting contract, or with all due respect, take your demands somewhere else.
I know. Discourse is an OSS project. To which I contributed many hours to the best of my abilities.
I didn’t make any demand. I said “please”. Maybe too many times. Please accept my apologies.
I raised the issue here:
Totally fine with a PR that removes all the cache_rows: true from the importers using mysql2 for now.
This appears to have been fixed: